SAP is extend­ing its embrace of gen­er­a­tive AI to the devel­op­er community. 

At SAP’s TechEd event host­ed last week in Ban­ga­lore, India, the com­pa­ny announced inno­va­tions for gen­er­a­tive AI appli­ca­tion devel­op­ment and new vec­tor data­base capa­bil­i­ties, along with AI-cen­tric learn­ing oppor­tu­ni­ties for developers. 

Chief among the announcements: 

• SAP Build Code: a pro-code appli­ca­tion devel­op­ment solu­tion lever­ag­ing SAP’s gen­er­a­tive AI copi­lot Joule, opti­mized for devel­op­ers using Java and JavaScript, that will enable SAP devel­op­ers to extend SAP S/4HANA as well as oth­er SAP and non-SAP sys­tems, as the next phase of SAP Build.
• A new vec­tor data­base for SAP HANA Cloud, mak­ing it eas­i­er for AI mod­els to man­age unstruc­tured data and enhanc­ing inter­ac­tions between large lan­guage mod­els (LLMs) and orga­ni­za­tions’ mis­sion-crit­i­cal data.
• The AI Foun­da­tion for SAP Busi­ness Tech­nol­o­gy Plat­form (BTP), described as ​“a new one-stop shop for devel­op­ers to cre­ate AI- and gen­er­a­tive AI-pow­ered exten­sions and appli­ca­tions on SAP BTP.”
• New ​“seman­tic onboard­ing” capa­bil­i­ties for SAP Data­s­phere, aimed at help­ing cus­tomers deliv­er data with busi­ness con­text and log­ic intact.
• New gen­er­a­tive AI capa­bil­i­ties and deep­er inte­gra­tion between SAP Build Process Automa­tion, SAP Sig­navio, and SAP Inte­gra­tion Suite, to accel­er­ate inte­gra­tion, automa­tion, and mon­i­tor­ing of busi­ness process­es in het­ero­ge­neous environments.
• New SAP Inte­gra­tion Suite Edge Inte­gra­tion Cell run­time, enabling hybrid deploy­ments and giv­ing cus­tomers in het­ero­ge­neous envi­ron­ments the flex­i­bil­i­ty to design inte­gra­tion flows in the cloud and then man­age them on-premises. 

Col­lec­tive­ly, the inno­va­tions will enable enter­pris­es to build their own appli­ca­tions and improve their pro­duc­tiv­i­ty in work­ing with gen­er­a­tive AI and large lan­guage mod­els — a major step for­ward for tech pro­fes­sion­als seek­ing to devel­op in inte­grat­ed hybrid cloud landscapes. 

“Today’s dynam­ic tech­nol­o­gy and busi­ness land­scape means every devel­op­er needs to be an AI devel­op­er,” said Juer­gen Mueller, Chief Tech­nol­o­gy Offi­cer and mem­ber of the Exec­u­tive Board of SAP SE, who’ll deliv­er the open­ing keynote at ASUG Tech Con­nect this week. 

“The inno­va­tions we’re launch­ing at SAP TechEd, from AI-infused pro-code tools to a one-stop shop to cre­ate gen­er­a­tive AI exten­sions and appli­ca­tions on SAP Busi­ness Tech­nol­o­gy Plat­form, sup­port the devel­op­ers at the heart of the AI rev­o­lu­tion and pro­vide them with resources they need to trans­form the way busi­ness­es run.” 

SAP Build Code

At last year’s TechEd, SAP unveiled SAP Build, uni­fy­ing its low-code appli­ca­tions in a suite of solu­tions aimed at cit­i­zen devel­op­ers.

Fol­low­ing on from that announce­ment, SAP this year announced SAP Build Code, aimed at all SAP devel­op­ers, which offers a set of AI-infused pro­duc­tiv­i­ty tools opti­mized for Java and JavaScript devel­op­ment. Along­side pre-built inte­gra­tions, APIs, con­nec­tors to SAP sys­tems and soft­ware from oth­er ven­dors, guid­ed tem­plates, and built-in SAP devel­op­er best prac­tices, the suite will embed Joule, SAP’s gen­er­a­tive-AI copi­lot, pro­vid­ing code gen­er­a­tion capa­bil­i­ties for data mod­el, appli­ca­tion log­ic, and test script creation. 

The new tools will also offer deep­er inter­op­er­abil­i­ty with SAP’s pro­pri­etary ABAP Cloud devel­op­ment mod­el and pro­gram­ming lan­guage through a side-by-side exten­si­bil­i­ty mod­el, enhanc­ing the abil­i­ty of pro-code devel­op­ers to col­lab­o­rate with low-code or no-code devel­op­ers work­ing in SAP Build. 

“Our phi­los­o­phy is that every­body can be a devel­op­er,” said JG Chi­ra­pu­rath, Chief Mar­ket­ing and Solu­tions Offi­cer at SAP, in an inter­view with ASUG. ​“If you look at SAP appli­ca­tions, we have appli­ca­tions for every room of the house — finance, HR, mar­ket­ing and sales, sup­ply chain — and to all of those users, our mes­sage is that they can pick up SAP Build and be pro­duc­tive. They don’t have to wait for cen­tral IT to assign them resources to get started.”

Bring­ing togeth­er pre­vi­ous­ly dis­parate design and run-time ser­vices, along­side a uni­fied appli­ca­tion life­cy­cle man­age­ment soft­ware with built-in gov­er­nance capa­bil­i­ties, the suite is meant to enhance col­lab­o­ra­tion between pro­fes­sion­al devel­op­ers, cit­i­zen devel­op­ers, and IT departments. 

“To cen­tral IT and hard­core devel­op­ers, we will bring mod­ern modal­i­ties,” Chi­ra­pu­rath con­tin­ued. ​“That’s where Java and JavaScript come in; and if you’ve grown up with ABAP, we’ll make that rel­e­vant in the con­text of all we’re doing. There are no islands here. It’s one land mass we’re con­nect­ing together.”

SAP Build Code will be made avail­able to select cus­tomers and part­ners ini­tial­ly before becom­ing gen­er­al­ly avail­able in Q1 2024. 

Vec­tor Data­base Fea­tures for SAP HANA Cloud

SAP also announced that it is enhanc­ing SAP HANA Cloud, its data­base man­age­ment sys­tem, to man­age unstruc­tured data (such as text, images, or audio) through a vec­tor datastore. 

“With this announce­ment, we get clos­er to the holy grail of gen­er­a­tive AI where you can com­bine the strength of large lan­guage mod­els with accu­rate real-time com­pa­ny data,” Mueller said on an ana­lyst call.

Retriev­ing data and con­text in the native lan­guage of LLMs, vec­tors (long arrays of num­bers that give val­ue to a large num­ber of attrib­ut­es or dimen­sions) will allow SAP HANA Cloud users to gain more con­text on their busi­ness data. With this func­tion­al­i­ty, the SAP HANA Cloud data­base can be lever­aged to con­tin­u­ous­ly train LLMs under­neath orga­ni­za­tions’ gen­er­a­tive AI appli­ca­tions with up-to-date, rel­e­vant busi­ness infor­ma­tion. This, accord­ing to SAP, will sim­pli­fy inter­ac­tions with LLMs and encour­age devel­op­ers to build gen­er­a­tive AI into their own applications. 

“When you train large lan­guage mod­els, it’s not enough to just train them with trans­ac­tion­al data or ana­lyt­i­cal data that’s sit­ting in a rela­tion­al data­base,” said Chi­ra­pu­rath. ​“You have to be able to train, store, and asso­ciate all types of data to train your LLM. Data­bas­es that can do that — vec­tor data­bas­es — employ a method called ​“vec­tor sim­i­lar­i­ty search,” which asks, if you search for a type of doc­u­ment, what else can you find inside this cor­pus of data that looks like it?”

The vec­tor search func­tion­al­i­ty, avail­able in Q1 2024, will also be avail­able in all three main hyper­scalers’ clouds, for those devel­op­ing out­side of SAP applications. 

AI Foun­da­tion

AI Foun­da­tion brings togeth­er var­i­ous tools to empow­er devel­op­ers to start build­ing busi­ness-ready AI appli­ca­tions and exten­sions. Pro­vid­ing ready-to-use AI ser­vices, access to LLMs, vec­tor data­base capa­bil­i­ties, and AI run­time and life­cy­cle man­age­ment, AI Foun­da­tion is anoth­er incen­tive for devel­op­ers to build their solu­tions on SAP BTP. 

Under the ban­ner of AI Foun­da­tion, SAP plans to offer AI ser­vices, gen­er­a­tive AI man­age­ment, AI work­load man­age­ment, access to busi­ness data and con­text, and life­cy­cle man­age­ment based on SAP HANA Cloud and dif­fer­ent SAP BTP fea­tures, such as the SAP AI Core run­time and SAP AI Launch­pad ser­vice for man­ag­ing AI use cas­es or scenarios.

AI Foun­da­tion will lat­er this quar­ter include a gen­er­a­tive AI hub with­in SAP AI Core, giv­ing devel­op­ers access to LLMs from dif­fer­ent providers to orches­trate mul­ti­ple mod­els and take advan­tage of tool­ing and test­ing capa­bil­i­ties. This hub will also con­nect to the vec­tor search capa­bil­i­ties with­in SAP HANA Cloud, allow­ing devel­op­ers to incor­po­rate con­tex­tu­al data into spe­cif­ic use cases.

‘Seman­tic Onboard­ing’ for SAP Datasphere

SAP rolled out new ​“seman­tic onboard­ing” capa­bil­i­ties for SAP Data­s­phere, aimed at help­ing cus­tomers deliv­er data with busi­ness con­text and log­ic intact. With these seman­tic onboard­ing capa­bil­i­ties, orga­ni­za­tions will be bet­ter equipped to build and deliv­er their own busi­ness-ready data prod­ucts. Expand­ed capa­bil­i­ties include:

• Import enti­ties: bring­ing in busi­ness con­text from source sys­tems such as SAP S/4HANA and SAP Busi­ness Ware­house, and auto­mat­i­cal­ly reusing seman­ti­cal def­i­n­i­tions and associations.
• Sim­pli­fied data shar­ing: access to cus­tomer and part­ner-man­aged data prod­ucts via APIs, in and out­side of SAP Datasphere.
• Pre­built busi­ness con­tent: includ­ing ready-to-use tem­plates, dash­boards, and reports that aim to de-risk data and ana­lyt­ics projects.

“SAP Data­s­phere is going to become a key ele­ment for part­ners and cus­tomers in terms of ratio­nal­iz­ing the data com­plex­i­ties they face,” said Joshua Green­baum, Prin­ci­pal at Enter­prise Appli­ca­tions Con­sult­ing. ​“Every com­pa­ny deals with this, regard­less of size, indus­try, or geog­ra­phy. They have major prob­lems with con­sol­i­dat­ing their dif­fer­ent kinds of data into process­es that can be lever­aged effectively.

“You can’t do any AI with­out clean, well-gov­erned data, and that applies for oper­a­tions and trans­ac­tion­al func­tions as well,” added Green­baum. ​“The focus on data qual­i­ty in SAP Data­s­phere, amid the inte­gra­tion of data that gen­er­a­tive AI is cre­at­ing, will have a pos­i­tive impact.”

SAP Inte­gra­tion Suite

AI-enabled process rec­om­men­da­tions, sim­pli­fied user expe­ri­ence, and eas­i­er-to-imple­ment process improve­ments are among gen­er­a­tive AI capa­bil­i­ties and inte­gra­tions announced between SAP Build Process Automa­tion solu­tions, SAP Sig­navio solu­tions, and SAP Inte­gra­tion Suite. 

Col­lec­tive­ly, these inno­va­tions will allow cus­tomers to bet­ter inte­grate, auto­mate, and mon­i­tor busi­ness process­es across SAP and non-SAP appli­ca­tions. SAP Sig­navio will now con­nect direct­ly to inte­gra­tion flows in SAP Inte­gra­tion Suite, allow­ing process ana­lysts to ana­lyze process per­for­mance across both SAP sys­tems and soft­ware from oth­er vendors.

With Joule, SAP Build Process Automa­tion can now assist users in cre­at­ing work­flows and busi­ness rules, while increased inter­op­er­abil­i­ty with SAP Build Apps and access to pre­built inte­gra­tions and automa­tions in SAP Busi­ness Accel­er­a­tor Hub will make it eas­i­er to devel­op automa­tions. Advanced event mesh capa­bil­i­ties for SAP Inte­gra­tion Suite can cap­ture com­mon busi­ness events, ini­ti­at­ing work­flows in SAP Build Process Automa­tion and draw­ing sup­port from process min­ing data from SAP Signavio. 

A new Edge Inte­gra­tion Cell run­time is now gen­er­al­ly avail­able for SAP Inte­gra­tion Suite, enabling improved inte­gra­tion and API man­age­ment in both cus­tomer-man­aged data cen­ters and pri­vate clouds. 

For those cus­tomers in het­ero­ge­neous envi­ron­ments, this will ensure greater flex­i­bil­i­ty and con­trol over inte­gra­tions and APIs while allow­ing them to deploy inte­gra­tion flows designed in the cloud in their on-premis­es envi­ron­ments; this hybrid deploy­ment option allows cus­tomers to keep cer­tain inte­gra­tions local while ben­e­fit­ting from cloud design capabilities. 

The run­time can sup­port run­ning inte­gra­tions in pri­vate land­scapes with­out rout­ing data to the cloud, as well as remote and offline inte­gra­tion sce­nar­ios (such as sys­tems in ware­hous­es or on ships), while main­tain­ing com­pli­ance with region-spe­cif­ic laws and regulations. 

“SAP is sell­ing itself short by only focus­ing on gen­er­a­tive-AI capa­bil­i­ties in recent announce­ments,” said Green­baum, who referred to the TechEd announce­ments as col­lec­tive­ly telling a ​“strong story.”

“With SAP Build Code and SAP Inte­gra­tion Suite, SAP is encour­ag­ing cus­tomers and part­ners to build and man­age next-gen­er­a­tion func­tion­al­i­ty across the board, whether we’re talk­ing about gen­er­a­tive AI or indus­try-spe­cif­ic activ­i­ties,” he added. ​“And SAP is clear­ly look­ing to sup­port devel­op­ment in inte­grat­ed hybrid cloud environments.”

Addi­tion­al Announcements

Two new knowl­edge resources — the SAP BTP Guid­ance Frame­work for design­ing, build­ing, and deploy­ing appli­ca­tions using SAP BTP, and a new developer’s guide for SAP BTP, focused on imple­ment­ing busi­ness appli­ca­tions on SAP BTP using ABAP Cloud and SAP Cloud Appli­ca­tion Pro­gram­ming Mod­el (CAP) — are now available. 

SAP Ana­lyt­ics Cloud, which already fea­tures a just ask fea­ture akin to Joule, allow­ing users to quick­ly receive data-dri­ven insights in response to their ques­tions, will launch plan­ning and ana­lyt­ics fea­tures to enhance data mod­el­ing and data inte­gra­tion, includ­ing a user-friend­ly data mod­el­ing expe­ri­ence and a new light­weight view­er to improve performance.

SAP also announced new role-based cer­ti­fi­ca­tion and free learn­ing resources aimed at upskilling devel­op­ers, par­tic­u­lar­ly around back-end devel­op­ers using the ABAP Cloud devel­op­ment model. 

SAP announced that it has joined the cor­po­rate affil­i­ate pro­gram at the Stan­ford Insti­tute for Human-Cen­tered Arti­fi­cial Intel­li­gence, con­nect­ing SAP researchers and engi­neers to the aca­d­e­m­ic com­mu­ni­ty to exam­ine the role of gen­er­a­tive AI in business.

And the com­pa­ny rolled out a new rewards pro­gram for RISE with SAP cus­tomers who run SAP S/4HANA Cloud, pri­vate edi­tion, with an ERP clean core. As part of the pro­gram, RISE cus­tomers who main­tain a clean core can earn annu­al cred­its through the SAP BTP Enter­prise Agree­ment to pay for SAP BTP usage.

For more on these announce­ments, read SAP’s offi­cial TechEd report.